Fix: Only set time-based features to NaN for rows with missing time index (closes #2700)Fix: Only set time-based features to NaN for rows with missing time i…

[realranjan]

 ## What does this PR do?
 - Fixes [Issue #2700](https://github.com/alteryx/featuretools/issues/2700): Previously, if any row in a DataFrame had a missing time index, all features for that row became NaN.
 - Now, only time-based features are set to NaN for rows with missing time index; non-time-based features are still computed.
 - Adds/updates tests (including `test_missing_time_index_rows`) to verify correct behavior.

 ## How was this tested?
 - All tests pass locally.
 - The new test ensures only time-based features are NaN for missing time index rows, while non-time-based features are always computed.

 ## Checklist
 - [x] Fix implemented
 - [x] Tests updated/added
 - [x] All tests passing
 - [x] Linked to #2700

 Closes #2700

[CLAassistant]

All committers have signed the CLA.